Biochemistry 1980, 19, 220-228
220
Fidelity of Fractionated Deoxyribonucleic Acid Polymerases from Human Placenta+ Sharon Wald Krauss and Stuart Linn*
ABSTRACT:Deoxyribonucleic acid (DNA) polymerase activities from human placenta have been fractionated and classified among the cy, p, and y types by using the criteria of size, ability to utilize various synthetic template-primers, and sensitivity to phosphate, N-ethylmaleimide, aphidicolin, and 2’,3’-dideoxythymidine 5’-triphosphate. Reverse transcriptase and &-polymeraseactivities were not detected. cy activity resolved into two fractions on diethylaminoethylcellulose which could be distinguished catalytically by their template-primer preferences. Similarly, multiple species of P-polymerase were obtained upon purification by phosphocellulose or glycerol gradient sedimentation. Each DNA polymerase fraction was tested for misincorporation frequencies with synthetic template-primers and either Mn2+or Mg2+. All a-polymerases copied the templates with relatively high fidelity, though ab-
-
%e
DNA polymerases of procaryotes have an intrinsic 3’ 5’ exonuclease activity which removes nucleotides that are incorrectly inserted during polymerization (Brutlag & Kornberg, 1972; Gefter, 1975). Error-free DNA synthesis also has been suggested to result from “kinetic proofreading” in the case of T4 phage polymerase (Hopfield, 1974; Gillin & Nossal, 1976a,b). However, the mechanism by which eucaryotes ensure a high degree of accuracy during DNA replication has not been identified yet. Neither partially purified nor homogeneous preparations of cy-, p-, or y-polymerases appear to have intrinsic 3’ 5’ exodeoxyribonuclease activity. In the case where a 3’ 5’ exonuclease activity has been reported to be associated with a polymerase (“6”; Byrnes et al., 1976), it has not been shown to be correction-specific. Moreover, Chang (1973) ruled out forms of editing which involve DNA hydrolysis by determining that nucleoside monophosphates are not generated during polymerization reactions with either the 3.4s or the 6.8s calf thymus polymerases. Therefore, it would appear that for eucaryotic DNA polymerases, either extrinsic factors act to assure fidelity or fidelity is controlled prior to phosphodiester bond formation. Recently, it was reported that DNA polymerase activity present in extracts from late-passage cultured MCR-5 human fibroblasts had levels of incorporation of incorrect nucleotides in the presence of a defined synthetic DNA template that were up to 10-fold higher than those observed with extracts from an early passage of these cells (Linn et al., 1976). These results are consistent with suggestions that alterations in the fidelity of DNA polymerase(s) could be important in aging or agerelated diseases (Szilard, 1959; Burnet, 1974). However, the preliminary observations must be further refined in order to determine whether the changes noted were due to primary or
--
From the Department of Biochemistry, University of California, Berkeley, California 94720. Received July 12, 1979. This research was supported by Contract EY-76-5-03-0034 from the Department of Energy, by Research Grant AGO0819 from the National Institutes of Health, and by Fellowship 5-F32-AG05084-01 from the National Institutes of Health to S.W.K. S.W.K. is a Senior Fellow of the California Division of the American Cancer Society (D-304).
0006-2960/80/0419-0220$01 .OO/O
solute values depended upon the length of primer and the primerltemplate ratio. With Mn2+present, an 8.6s species of P-polymerase copied the synthetic templates very accurately, but 6.3s and 4.6s forms copied the same polymers relatively unfaithfully. The 8.6s form could be converted to a 4.8s species with a concomitant loss of fidelity. y-Polymerase also copied deoxyribopolymers relatively inaccurately in the presence of Mn2+. The relative concentrations of complementary to noncomplementary triphosphates affected the frequency of misincorporation for 3/ and y activities. Conversely, a-polymerase did not show similar behavior at saturating triphosphate concentrations. None of the placental polymerase fractions contained exonuclease activity that could discriminate between complementary and noncomplementary 3’ ends.
secondary structural changes in the polymerases, to posttranslational modifications of normal polymerases, to the appearance of a new, error-prone polymerase species, or to an alteration of other polypeptide components that are involved in assuring fidelity during DNA replication. In this context, we have isolated DNA polymerases from human placenta and examined the accuracy with which they copy synthetic templates. As reported here, under comparable assay conditions, a-polymerases copied a variety of synthetic templates with relatively high fidelity, whereas p- and y-polymerases copied the same templates less faithfully. However, with @-polymerase,the particular form of the enzyme has a profound effect on its fidelity. In addition, the fidelities of the polymerases have varying sensitivities to the relative concentrations of complementary and noncomplementary triphosphates, the primer lengths, and the primerltemplate ratios. With this information and the observation that the polymerase fractionation applies also to cultured fibroblasts, we should be able to study more completely the observations made with the aging cells. Experimental Procedures Materials. DEAE-cellulose, type 40, was obtained from Brown Co., Berlin, N H ; phosphocellulose (P11) was from Whatman. All synthetic polynucleotides and oligonucleotides were from P-L Biochemicals. [ 3H]Deoxyribonucleoside triphosphates were from Amersham, and [cd2P]dATPwas from New England Nuclear. Unlabeled deoxyribonucleoside triphosphates were from P-L Biochemicals. Salmon sperm DNA was activated according to the procedure of Schlabach et al. (1971). Markers for glycerol gradient sedimentation were bacterial alkaline phosphatase (Worthington) and ovalbumin (Sigma). Homogeneous Escherichia coli polymerase I was the generous gift of Dr. Arthur Kornberg, Stanford University. Terminal deoxyribonucleotidyltransferasewas kindly provided by Dr. Robert Ratliff, Los Alamos, NM. P~ly(dA-[~HldT) was synthesized according to Modrich & Lehman (1970). The final product had a specific activity of 23 000 cpm/nmol. [ L Y - ~ ~ P I ~was G Tadded P to 3’ termini of
0 1980 American Chemical Society
HUMAN DNA POLYMERASES
poly(dA-[3H]dT) with terminal transferase in a 50-pL reaction which contained 0.2 M potassium cacodylate, pH 7.1, 1 mM CoCl2, 1 mM 2-mercaptoethanol, 30 nmol of ~ o l y ( d A - [ ~ H ] dT), and 13 pM [ E - ~ ~ P I ~ G T The P . number of residues added per molecule was calculated after determining the number of 5' termini of the p ~ l y ( d A - [ ~ H ] d Twith ) the polynucleotide kinase exchange reaction (Berkner & Folk, 1977). The polymer was extracted with phenol and dialyzed. Poly(dAdT)-[3H]dT and p~ly(dA-dT)-[~H]dG were synthesized by the addition of the nucleoside triphosphate to commercial poly(dA-dT). [3H]dGTP was added by using terminal transferase in a 0.2-mL reaction mixture containing 0.2 M potassium cacodylate, pH 7.1, 1 mM CoCl2, 1 mM 2mercaptoethanol, 55 nmol of poly(dA-dT), and 10 pM [3H]dGTP (12.6 Ci/mmol). [3H]dTTP was added by using E . coli DNA polymerase I in a 0.2-mL reaction containing 80 mM Tris-HC1, pH 7.5, 6 mM MgC12, 1 mM 2mercaptoethanol, 50 nmol of poly(dA-dT), and 20 pM [3H]dTTP (30 Ci/mmol). All polymers were extracted with phenol and dialyzed extensively before use. Buffer I contained 0.4 M potassium phosphate, pH 7.5,0.5 mM dithiothreitol, and 10 mM 2-mercaptoethanol. Buffer I1 contained 0.02 M potassium phosphate, pH 7.5, 0.5 mM dithiothreitol, and 1 mM 2-mercaptoethanol. Buffer I11 contained 0.05 M potassium phosphate, pH 7.5, 0.5 mM dithiothreitol, and 1 mM 2-mercaptoethanol. Buffer IV contained 0.05 M potassium phosphate, pH 7.5, 0.5 mM dithiothreitol, 1 mM 2-mercaptoethanol, and 20% glycerol. Polymerase Assay. Enzyme activity was assayed by using activated salmon sperm DNA under the conditions described by Linn et al. (1976). When measuring a-polymerase activity, KCl was omitted. One unit of enzyme incorporates 1 nmol of total nucleotide in 30 min at 37 "C. Fidelity Assay. Reaction mixtures (0.1 mL) contained 50 mM Tris-HC1, pH 7.5, 0.5 mM dithiothreitol, 0.5 mg/mL bovine serum albumin, 10 nmol of synthetic templateprimer, 50 pM complementary deoxyribonucleoside triphosphate, 6 pM noncomplementary deoxyribonucleoside triphosphate, and enzyme as indicated. Initally, noncomplementary triphosphates were purified as described by Lehman et al. (1958) and tested in misincorporation assays with homogeneous E . coli DNA polymerase I and placental polymerases. Subsequently, unpurified triphosphates were used in misincorporation assays only if there were no detectable differences in error frequencies relative to purified triphosphates. (The controls with the E. coli enzyme also served to assure that the polymers were free of contaminating nucleotides.) Unless otherwise noted, assays of a-polymerases also contained 0.25 mM MnC12 and no KC1, assays of 0-polymerases contained 0.5 mM MnC12 and 50 mM KC1, and assays of y-polymerase contained 0.5 mM MnC12 and 100 mM KC1. In assays with Mg2+instead of Mn2+, MgC12 was 1 mM with poly(dA).poly(dT), poly(dA-dT), or p o l y ( d T ) ~ ~ l i g o ( d A )2~ ~mM - ~ ~with , poly(d1). poly(dC) or poly(d1-dC), and 7.5 mM with poly(rA)-oligo(dT) Before the assay, template-primers were heated for 10 min at 70 "C and slowly cooled to assure reproducible secondary structures. Each misincorporation ratio was obtained from parallel reactions, one with complementary [3H]deoxyribonucleosidetriphosphate (80-1 50 cpm/pmol) to measure total polymer synthesis and the other with noncomplementary [3H]deoxyribonucleoside triphosphate (1 2-30 cpm/fmol) to measure nonfaithful synthesis. Blank values for misincorporated nucleotide were reproducible to within 20% in an individual experiment, and values 50% above the blank were considered significant. Each enzyme sample was tested
V O L . 19, N O .
1,
1980
221
for contaminating endogenous template with controls in which templateprimer was omitted and for terminal transferase with controls using single-stranded homopolymer. After incubation for 4 h at 37 "C, reaction mixtures were precipitated and collected onto glass fiber filters and radioactivity was measured as described by Linn et al. (1976). Synthetic templateprimer concentrations are expressed as nucleotide residues. Reactions were generally done with several levels of enzyme to assure a constant misincorporation frequency vs. the level of synthesis. For brevity, however, only the highest levels of reaction are noted in the tables. 3'- 5'Exonuclease Assay. Reaction mixtures (0.1 mL) contained 80 mM Hepes,' pH 7.0, 1 mM MnC12, 120 mM KCl, 0.5 mM dithiothreitol, enzyme as indicated, and either 10 nmol of p~ly(dA-[~H]dT)-[~~P]dG, 0.5 nmol of poly(dAdT)-[3H]d(T)o,2,or 0.7 nmol of pol~(dA-dT)-[~H]d(G)~~. At various times after incubation at 37 "C, 20-pL aliquots were precipitated onto glass fiber filters and radioactivity was determined as described by Linn et al. (1976). Protein Assay. Protein was determined according to Lowry et al. (1951) with bovine serum albumin as a standard. Fractionation of Placental Polymerases. All operations were at 0-4 "C. In a typical preparation (Table I) a human placenta was immersed in ice-cold 25 mM potassium phosphate, pH 7.0, 0.15 M NaCl, and 0.015 M sodium citrate within 10 min of delivery by Caesarian section. Approximately 400 g of tissue was excised, rinsed, suspended in 250 mL of buffer I, and homogenized in a Waring blender 4 times for 30 s. The homogenate was centrifuged for 30 min at 13000g, and the supernatant was saved. The pellet was resuspended in 300 mL of buffer I and blended as previously described, and then this suspension was sonicated for four, 30-s intervals by using the large probe of a Branson sonifier cell disrupter (Model W185D) at 75 W. The sonicate was centrifuged as above, and the supernatant was combined with the first supernatant. The crude extract was passed at 2 mL/min over a column of DEAE-cellulose (5.5 X 27.5 cm) previously equilibrated with buffer I. Fractions containing significant red hemoglobin color were pooled, concentrated threefold by dialysis against 30% polyethylene glycol (w/v), 0.05 M potassium phosphate, pH 7.5,0.5 mM dithiothreitol, and 10 mM 2-mercaptoethanol, and then dialyzed extensively against buffer 11. The dialyzed extract was loaded onto a second DEAEcellulose column (4 X 36 cm) which had been equilibrated with buffer 11. The column was washed with 350 mL of this buffer, and fractions containing the red hemoglobin were pooled. A 2-L linear gradient of potassium phosphate, pH 7.5, from 0.02 to 0.40 M containing 0.5 mM dithiothreitol and 1 mM 2mercaptoethanol was finally passed through the column, and fractions of 12 mL were collected into plastic tubes. Four peaks of polymerase activity were recovered that eluted in the flow through and at 0.10,0.19, and 0.23 M phosphate (Figure 1). These were subsequently shown to be O-, y-, and two a-polymerases (called aIand a*),respectively (see below). Each pool of polymerase activity was concentrated two- to threefold by dialysis against 30% polyethylene glycol (w/v in buffer IV), dialyzed extensively against buffer 111, and stored at -80 "C. For further purification, individual polymerase pools were rapidly thawed, diluted with an equal volume of buffer IV, and adsorbed onto phosphocellulose columns (10 mg of protein I Abbreviation used: Hepes, 4-(2-hydroxyethyl)-l -piperazineethanesulfonic acid.
222
BIOCHEMISTRY
_ _ _ _ I
KRAUSS AND LlNN . _ _ _ _ . I _ _ I
Table 1: Purification of Polymerase Activitiesa -__ fraction
-____
-__.__I-__
_ I _
(1) (11)
(m)
crude extract first DEAE-cellulose dialyzed concentrate second DEAE-cellulose fi 1 a2
(IV)
protein (mg)
act. (units)
. I _ _ . -
_
.
-__I__
_
_
--._______I
sp act. (units/mgj
. l l . l l _ -
730 620 200
10300
132 140 100 60
1190 630 190 46
6440
4510 1090 1000
0.26
494
0.41 0.28
0.15
178 97 (400) 94 (850)
0.51 (2.1) 2.0 (18)
phosphocellulose 24 23 2.4 2.4 11 5.9
81
Y ala OL2a “2b
(VI
_____l___l_l__
vol (mL)
__
_.
33 50 3.6 2.6 2.3
13.1 130 17 ( 2 0 ) 15 (26) 15 (133) 5.9 ( 5 9 )
1.2
0.39 2.6 4.7 (13) 5.6 (23)
6.5 (59) 5.0 (50)
glycerol gradients Pi
0.16
Y
0.26 0.26 0.26 0.26 0.26
